The Number Is Not the Explanation
Revenue fell 18%. CPA increased 12%. Conversion rate improved. Brand search was down. Someone changed the bid strategy on Tuesday and performance recovered by Friday. Put enough of those numbers on a dashboard and it starts to feel like you understand what happened.
You might not.
A metric is an observation. An explanation needs more: what changed, when it changed, what else was happening, what people expected to happen and what evidence supports the connection. This is the boring but important difference between correlation and causation. The UK Government Analysis Function puts it plainly: showing that two things moved together is not enough to show that one caused the other. Establishing causation means showing the relationship, getting the sequence right and ruling out other plausible causes. In real marketing accounts, that last part is often the hard one.
The dashboard can tell you that something moved. It can’t tell you, on its own, why it moved.
This is where reporting often gets quietly upgraded into storytelling. Spend went down, ROAS went up, therefore cutting spend improved efficiency. Traffic fell after a site migration, therefore the migration caused the decline. Conversions recovered after generic search was paused, therefore generics were the problem. Any of those explanations might be right. But the chart itself doesn’t prove them. You still need the surrounding evidence: implementation dates, tracking changes, campaign structure, seasonality, other interventions, previous tests and sometimes the uncomfortable possibility that the answer is simply “we don’t know”.
Dashboards are useful because they compress complexity. That is also their danger. Once the context disappears, a number can survive for years while the explanation attached to it gradually hardens into fact. The answer isn’t more charts. It’s keeping the numbers connected to the decisions, changes and evidence that give them meaning.